Client caseE-mobility retail · 15 websites, 130+ models
1,486 chats a month across 15 websites and three messaging apps, with no rep on the front line
A text AI agent knows the full catalog, matches a model to the customer's needs, sends links and passes a ready lead to Bitrix24.
Project overview
- Industry
- E-mobility retail
- CRM
- Bitrix24
- Task
- Handling inbound inquiries from every channel
- Tool
- Text AI agent
- Channels
- Widget on 15 websites, Instagram, WhatsApp, Telegram
- Time to launch
- 12 days
- chats in a month
- 1,486
- of chats end in a lead
- 43%
- leads in a month
- 639
- missed inquiries (was 34%)
- 0%
Context
Where it started
The company runs 15 websites plus Instagram, WhatsApp and Telegram, with more than 130 models in the catalog. Inquiries come from every channel at once, and almost every one needs a recommendation: the person wants a vehicle for a purpose, not a specific SKU.
Reps could not keep up. The client's CRM data for the year showed that 34% of inquiries were missed: answered too late or not at all. A person who gets no reply on a website opens the next search tab.
Process
How the work went
Loading the catalog
The agent learned the full catalog: 130+ models, specs, prices and the differences between them. It knows what suits a child and what suits a courier.
Recommendation and qualification script
The agent finds out what the vehicle is for and picks models based on the answers:
- for delivery work
- for a child
- for city commuting
- what top speed is needed
It sends matching models with product links and answers questions about them.
A custom widget for 15 websites
We built a chat widget and installed it on all 15 of the client's websites. One agent works across every site.
Connecting Instagram, WhatsApp and Telegram
The same agent answers on Instagram and in messaging apps, with the same knowledge base and rules as on the websites.
Bitrix24 integration
The lead goes to Bitrix24 and the rep gets a notification. Qualification happens without a person.
Launch
12 days from brief to launch.
For the customer
What it looks like for the company's customer
- 01A person writes in the website widget: “I need a scooter for getting around the city.”
- 02The agent asks for details, such as the top speed they need, and sends matching models with links.
- 03It answers questions about the models.
- 04It creates the lead: the lead lands in Bitrix24, and the rep gets a notification.
Result
After one month
| Stage | Count | Conversion |
|---|---|---|
| Chats | 1,486 | 100% |
| Leads | 639 | 43% |
| Metric | Before | After |
|---|---|---|
| Missed inquiry rate | 34% | 0%: the agent handles every inquiry |
| Channels per handler | — | 18: 15 websites + Instagram + WhatsApp + Telegram |
| Lead qualification | rep | agent, no person involved |
| Model selection | rep | agent, from 130+ products |
Outcome
What the client got
- leads in a month, collected without reps
- 639
- channels served by one agent
- 18
- missed inquiries
- 0%
The main point
One agent covers 15 websites, Instagram, WhatsApp and Telegram at once and knows the whole catalog. The company did not have to hire reps for each channel, no inquiry goes unanswered, and reps get leads that are already qualified.
Related
- Industry case studiesAI agent case studies in ecommerceHow a text AI agent advises shoppers on websites and in messaging apps, picks products from the catalog and sends ready leads to the CRM, in numbers.Learn more
- Solution for the industryAI chatbotAnswers by SMS, WhatsApp, Instagram and website chat from your documents.Learn more
- SolutionBitrix24 integrationThe AI agent writes deals, tasks and call and chat summaries to Bitrix24.Learn more